Refine
H-BRS Bibliography
- yes (41) (remove)
Departments, institutes and facilities
- Fachbereich Informatik (41) (remove)
Document Type
- Conference Object (28)
- Article (10)
- Report (2)
- Part of a Book (1)
Year of publication
- 2018 (41) (remove)
Has Fulltext
- no (41) (remove)
Keywords
- Surrogate Modeling (2)
- Virtual Reality (2)
- 3D User Interface (1)
- 3D user interface (1)
- Alternatives (1)
- Available Bandwidth (1)
- B2T (1)
- Bandwidth Estimation (1)
- Bayesian Network (1)
- Bayessches Netz (1)
In presence of conflicting or ambiguous visual cues in complex scenes, performing 3D selection and manipulation tasks can be challenging. To improve motor planning and coordination, we explore audio-tactile cues to inform the user about the presence of objects in hand proximity, e.g., to avoid unwanted object penetrations. We do so through a novel glove-based tactile interface, enhanced by audio cues. Through two user studies, we illustrate that proximity guidance cues improve spatial awareness, hand motions, and collision avoidance behaviors, and show how proximity cues in combination with collision and friction cues can significantly improve performance.
We present a novel forearm-and-glove tactile interface that can enhance 3D interaction by guiding hand motor planning and coordination. In particular, we aim to improve hand motion and pose actions related to selection and manipulation tasks. Through our user studies, we illustrate how tactile patterns can guide the user, by triggering hand pose and motion changes, for example to grasp (select) and manipulate (move) an object. We discuss the potential and limitations of the interface, and outline future work.
Entering the work envelope of an industrial robot can lead to severe injury from collisions with moving parts of the system. Conventional safety mechanisms therefore mostly restrict access to the robot using physical barriers such as walls and fences or non-contact protective devices including light curtains and laser scanners. As none of these mechanisms applies to human-robot-collaboration (HRC), a concept in which human and machine complement one another by working hand in hand, there is a rising need for safe and reliable detection of human body parts amidst background clutter. For this application camera-based systems are typically well suited. Still, safety concerns remain, owing to possible detection failures caused by environmental occlusion, extraneous light or other adverse imaging conditions. While ultrasonic proximity sensing can provide physical diversity to the system, it does not yet allow to reliably distinguish relevant objects from background objects.This work investigates a new approach to detecting relevant objects and human body parts based on acoustic holography. The approach is experimentally validated using a low-cost application-specific ultrasonic sensor system created from micro-electromechanical systems (MEMS). The presented results show that this system far outperforms conventional proximity sensors in terms of lateral imaging resolution and thus allows for more intelligent muting processes without compromising the safety of people working close to the robot. Based upon this work, a next step could be the development of a multimodal sensor systems to safeguard workers who collaborate with robots using the described ultrasonic sensor system.
Almost unnoticed by the e-learning community, the underlying technology of the WWW is undergoing massive technological changes on all levels these days. In this paper we draw the attention to the emerging game changer and discuss the consequences for online learning. In our e-learning project "Work & Study", funded by the German Federal Ministry of Education and Research, we have experimented with several new technological approaches such as Mobile First, Responsive Design, Mobile Apps, Web Components, Client-side Components, Progressive Web Apps, Course Apps, e-books, and web sockets for real time collaboration and report about the results and consequences for online learning practice. The modular web is emerging where e-learning units are composed from and delivered by universally embeddable web components.
In this paper we propose an architecture to integrate classical planning and real autonomous mobile robots. We start by providing with a high level description of all necessary components to set the goals, generate plans and execute them on real robots and monitor the outcome of their actions. At the core of our method and to deal with execution issues we code the agent actions with automatas. We prove the flexibility of the system by testing on two different domains: industrial (Basic Transportation Test) and domestic (General Purpose Service Robot) in the context of the international RoboCup competition. Additionally we benchmark the scalability of the planning system in two domains on a set of planning problems with increasing complexity. The proposed framework is open source1 and can be easily extended.
In recent years, a variety of methods have been introduced to exploit the decrease in visual acuity of peripheral vision, known as foveated rendering. As more and more computationally involved shading is requested and display resolutions increase, maintaining low latencies is challenging when rendering in a virtual reality context. Here, foveated rendering is a promising approach for reducing the number of shaded samples. However, besides the reduction of the visual acuity, the eye is an optical system, filtering radiance through lenses. The lenses create depth-of-field (DoF) effects when accommodated to objects at varying distances. The central idea of this article is to exploit these effects as a filtering method to conceal rendering artifacts. To showcase the potential of such filters, we present a foveated rendering system, tightly integrated with a gaze-contingent DoF filter. Besides presenting benchmarks of the DoF and rendering pipeline, we carried out a perceptual study, showing that rendering quality is rated almost on par with full rendering when using DoF in our foveated mode, while shaded samples are reduced by more than 69%.
General Chair Message
(2018)
3D user interfaces for virtual reality and games: 3D selection, manipulation, and spatial navigation
(2018)
In this course, we will take a detailed look at different topics in the field of 3D user interfaces (3DUIs) for Virtual Reality and Gaming. With the advent of Augmented and Virtual Reality in numerous application areas, the need and interest in more effective interfaces becomes prevalent, among others driven forward by improved technologies, increasing application complexity and user experience requirements. Within this course, we highlight key issues in the design of diverse 3DUIs by looking closely into both simple and advanced 3D selection/manipulation and spatial navigation interface design topics. These topics are highly relevant, as they form the basis for most 3DUI-driven application, yet also can cause major issues (performance, usability, experience. motion sickness) when not designed properly as they can be difficult to handle. Within this course, we build on top of a general understanding of 3DUIs to discuss typical pitfalls by looking closely at theoretical and practical aspects of selection, manipulation, and navigation and highlight guidelines for their use.
For many different applications, current information about the bandwidth-related metrics of the utilized connection is very useful as they directly impact the performance of throughput sensitive applications such as streaming servers, IPTV and VoIP applications. In literature, several tools have been proposed to estimate major bandwidth-related metrics such as capacity, available bandwidth and achievable throughput. The vast majority of these tools fall into one of Packet Pair (PP), Variable Packet Size (VPS), Self-Loading of Periodic Streams (SLoPS) or Throughput approaches. In this study, seven popular bandwidth estimation tools including nettimer, pathrate, pathchar, pchar, clink, pathload and iperf belonging to these four well-known estimation techniques are presented and experimentally evaluated in a controlled testbed environment. Differently from the rest of studies in literature, all tools have been uniformly classified and evaluated according to an objective and sophisticated classification and evaluation scheme. The performance comparison of the tools incorporates not only the estimation accuracy but also the probing time and overhead caused.
Large, high-resolution displays are highly suitable for creation of digital environments for co-located collaborative task solving. Yet, placing multiple users in a shared environment may increase the risk of interferences, thus causing mental discomfort and decreasing efficiency of the team. To mitigate interferences coordination strategies and techniques were introduced. However, in a mixed-focus collaboration scenarios users switch now and again between loosely and tightly collaboration, therefore different coordination techniques might be required depending on the current collaboration state of team members. For that, systems have to be able to recognize collaboration states as well as transitions between them to ensure a proper adjustment of the coordination strategy. Previous studies on group behavior during collaboration in front of large displays investigated solely collaborative coupling states, not transitions between them though. To address this gap, we conducted a study with 12 participant dyads in front of a tiled display and let them solve two tasks in two different conditions (focus and overview). We looked into group dynamics and categorized transitions by means of changes in proximity, verbal communication, visual attention, visual interface, and gestures. The findings can be valuable for user interface design and development of group behavior models.
Design optimization techniques are often used at the beginning of the design process to explore the space of possible designs. In these domains illumination algorithms, such as MAP-Elites, are promising alternatives to classic optimization algorithms because they produce diverse, high-quality solutions in a single run, instead of only a single near-optimal solution. Unfortunately, these algorithms currently require a large number of function evaluations, limiting their applicability. In this article we introduce a new illumination algorithm, Surrogate-Assisted Illumination (SAIL), that leverages surrogate modeling techniques to create a map of the design space according to user-defined features while minimizing the number of fitness evaluations. On a two-dimensional airfoil optimization problem SAIL produces hundreds of diverse but high-performing designs with several orders of magnitude fewer evaluations than MAP-Elites or CMA-ES. We demonstrate that SAIL is also capable of producing maps of high-performing designs in realistic three-dimensional aerodynamic tasks with an accurate flow simulation. Data-efficient design exploration with SAIL can help designers understand what is possible, beyond what is optimal, by considering more than pure objective-based optimization.
The increasing complexity of tasks that are required to be executed by robots demands higher reliability of robotic platforms. For this, it is crucial for robot developers to consider fault diagnosis. In this study, a general non-intrusive fault diagnosis system for robotic platforms is proposed. A mini-PC is non-intrusively attached to a robot that is used to detect and diagnose faults. The health data and diagnosis produced by the mini-PC is then standardized and transmitted to a remote-PC. A storage device is also attached to the mini-PC for data logging of health data in case of loss of communication with the remote-PC. In this study, a hybrid fault diagnosis method is compared to consistency-based diagnosis (CBD), and CBD is selected to be deployed on the system. The proposed system is modular and can be deployed on different robotic platforms with minimum setup.
Robot deployment in realistic dynamic environments is a challenging problem despite the fact that robots can be quite skilled at a large number of isolated tasks. One reason for this is that robots are rarely equipped with powerful introspection capabilities, which means that they cannot always deal with failures in a reasonable manner; in addition, manual diagnosis is often a tedious task that requires technicians to have a considerable set of robotics skills.
Robot deployment in realistic environments is challenging despite the fact that robots can be quite skilled at a large number of isolated tasks. One reason for this is that robots are rarely equipped with powerful introspection capabilities, which means that they cannot always deal with failures in an acceptable manner; in addition, manual diagnosis is often a tedious task that requires technicians to have a considerable set of robotics skills. In this paper, we discuss our ongoing efforts to address some of these problems. In particular, we (i) present our early efforts at developing a robotic black box and consider some factors that complicate its design, (ii) explain our component and system monitoring concept, and (iii) describe the necessity for remote monitoring and experimentation as well as our initial attempts at performing those. Our preliminary work opens a range of promising directions for making robots more usable and reliable in practice.
Surrogate-assistance approaches have long been used in computationally expensive domains to improve the data-efficiency of optimization algorithms. Neuroevolution, however, has so far resisted the application of these techniques because it requires the surrogate model to make fitness predictions based on variable topologies, instead of a vector of parameters. Our main insight is that we can sidestep this problem by using kernel-based surrogate models, which require only the definition of a distance measure between individuals. Our second insight is that the well-established Neuroevolution of Augmenting Topologies (NEAT) algorithm provides a computationally efficient distance measure between dissimilar networks in the form of "compatibility distance", initially designed to maintain topological diversity. Combining these two ideas, we introduce a surrogate-assisted neuroevolution algorithm that combines NEAT and a surrogate model built using a compatibility distance kernel. We demonstrate the data-efficiency of this new algorithm on the low dimensional cart-pole swing-up problem, as well as the higher dimensional half-cheetah running task. In both tasks the surrogate-assisted variant achieves the same or better results with several times fewer function evaluations as the original NEAT.
Improving the Performance of Parallel SpMV Operations on NUMA Systems with Adaptive Load Balancing
(2018)
For a parallel Sparse Matrix Vector Multiply (SpMV) on a multiprocessor, rather simple and efficient work distributions often produce good results. In cases where this is not true, adaptive load balancing can improve the balance and performance. This paper introduces a low overhead framework for adaptive load balancing of parallel SpMV operations. It uses statistical filters to gather relevant runtime performance data and detects an imbalance situation. Three different algorithms were compared that adaptively balance the load with high quality and low overhead. Results show that for sparse matrices, where the adaptive load balancing was enabled, an average speedup of 1.15 (regarding the total execution time) could be achieved with our best algorithm over 4 different matrix formats and two different NUMA systems.
Large, high-resolution displays demonstrated their effectiveness in lab settings for cognitively demanding tasks in single user and collaborative scenarios. The effectiveness is mostly reached through inherent displays' properties - large display real estate and high resolution - that allow for visualization of complex datasets, and support of group work and embodied interaction. To raise users' efficiency, however, more sophisticated user support in the form of advanced user interfaces might be needed. For that we need profound understanding of how large, tiled displays impact users work and behavior. We need to extract behavioral patterns for different tasks and data types. This paper reports on study results of how users, while working collaboratively, process spatially fixed items on large, tiled displays. The results revealed a recurrent pattern showing that users prefer to process documents column wise rather than row wise or erratic.
Preleukemic clones carrying BCR-ABLp190 oncogenic lesions are found in neonatal cord blood, where the majority of preleukemic carriers do not convert into precursor B-cell acute lymphoblastic leukemia (pB-ALL). However, the critical question of how these preleukemic cells transform into pB-ALL remains undefined. Here we model a BCR-ABLp190 preleukemic state and show that limiting BCR-ABLp190 expression to hematopoietic stem/progenitor cells (HS/PC) in mice (Sca1-BCR-ABLp190) causes pB-ALL at low penetrance, which resembles the human disease. pB-ALL blast cells were BCR-ABL-negative and transcriptionally similar to pro-B/pre-B cells, suggesting disease onset upon reduced Pax5 functionality. Consistent with this, double Sca1-BCR-ABLp190+Pax5+/- mice developed pB-ALL with shorter latencies, 90% incidence, and accumulation of genomic alterations in the remaining wild-type Pax5 allele. Mechanistically, the Pax5-deficient leukemic pro-B cells exhibited a metabolic switch towards increased glucose utilization and energy metabolism. Transcriptome analysis revealed that metabolic genes (IDH1, G6PC3, GAPDH, PGK1, MYC, ENO1, ACO1) were upregulated in Pax5-deficient leukemic cells, and a similar metabolic signature could be observed in human leukemia. Our studies unveil the first in vivo evidence that the combination between Sca1-BCR-ABLp190 and metabolic reprogramming imposed by reduced Pax5 expression is sufficient for pB-ALL development. These findings might help to prevent conversion of BCR-ABLp190 preleukemic cells.
Speech understanding is a fundamental feature for many applications focused on human-robot interaction. Although many techniques and several services for speech recognition and natural language understanding have been developed in the last years, specific implementation and validation on domestic service robots have not been performed. In this paper, we describe the implementation and the results of a functional benchmark for speech understanding in service robotics that has been developed and tested in the context of different robot competitions: RoboCup@Home, RoCKIn@Home and within the European Robotics League on Service Robots. Different approaches used by the teams in the competitions are presented and the evaluation results obtained in the competitions are discussed.